Vascular ageing (VA), characterised by arterial stiffness (AS) and endothelial dysfunction, is a major contributor to cardiovascular disease (CVD), yet its early detection is challenging. Emerging evidence suggests that gut microbiota (GM) influences VA, but studies in diverse populations remain scarce. To assess the association between GM composition and VA markers, focusing on sex- and ethnicity-specific differences. This cross-sectional study included 2,532 participants from the multi-ethnic HEalthy Life In an Urban Setting (HELIUS) cohort. Microbiome composition was assessed using 16S rRNA sequencing of stool samples. VA markers (aortic pulse wave velocity (AoPWV), central systolic blood pressure (cSBP), and augmentation index (AIx)) were measured using validated techniques. Alpha diversity was evaluated using the Shannon index, while beta diversity differences between groups were assessed with the PERMANOVA test on Bray-Curtis distances. Analysis of Compositions of Microbiomes with Bias Correction (ANCOM-BC) was used to identify differentially abundant genera, which were analysed in adjusted linear and logistic regression models, including sex and ethnicity interactions. VA markers significantly differed by ethnicity, with Dutch participants showing the most favourable profiles, while Surinamese and Ghanaian groups exhibited the worst. Beta diversity differed across all VA markers (p <0.001), while alpha diversity was only significant for cSBP (p <0.001). At the genus level, Faecalibacterium, Coprococcus 3, and Ruminococcaceae UCG-013 abundance were linearly associated with lower AoPWV (Figure 1), while Streptococcus increased AoPWV and AIx. Anaerostipes showed a protective effect on AoPWV only in men, while Dorea was associated with lower cSBP only in women. In logistic models (Figure 2), Bacteroides and Faecalibacterium reduced the odds of AS by 10% (95%CI: 0.82,0.99) and 11% (95%CI: 0.81,0.99), respectively, while Streptococcus increased the odds of AS by 8% (95%CI: 1.01,1.16). Lachnospiraceae UCG-010, Marvinbryantia, and Ruminococcaceae UCG-003 were protective against high cSBP (OR=0.86, 95%CI: 0.77,0.95; OR=0.87, 95%CI: 0.77,0.99; OR=0.82, 95%CI: 0.70,0.95, respectively). Notably, Lachnospiraceae UCG-010 and Ruminococcaceae UCG-013 were protective against AS and high cSBP in men. Ethnic-specific effects were observed, with Ruminococcaceae UCG-013 reducing AS and AIx in South-Asian Surinamese and high cSBP in Moroccans. GM composition is associated with VA, with notable differences across sex and ethnic groups. Our findings underscore the need for personalised approaches when considering microbiome-based interventions in diverse populations. Longitudinal studies are needed to clarify causal relationships and assess the clinical potential of microbiome-targeted interventions.Linear associations between taxa and VA Logistic associations - taxa and VA

Apathy is common after stroke and has been associated with cognitive impairment. However, causality between post-stroke apathy and cognitive impairment remains unclear. We assessed the course of apathy in relation to changes in cognitive functioning in stroke survivors. Using the Apathy Scale (AS) and cognitive tests on memory, processing speed and executive functioning at six- and 15 months post-stroke we tested for associations between (1) AS-scores and (change in) cognitive scores; (2) apathy course (persistent/incident/resolved) and cognitive change scores. Of 117 included participants, 29% had persistent apathy, 13% apathy resolving over time and 10% apathy emerging between 6-15 months post-stroke. Higher AS-scores were cross-sectionally and longitudinally associated with lower cognitive scores. Relations between apathy and cognitive change scores were ambiguous. These inconsistent relations between apathy and changes in cognition over time suggest that post-stroke apathy does not directly impact cognitive performance. Both these sequelae of stroke require separate attention.

Background: Hypertension can be classified into different phenotypes based on systolic and diastolic blood pressure (BP) that carry a different prognosis and may therefore be differently associated with sympathetic activity. We assessed the association between cardiac autonomic function determined from continuous finger BP recordings and hypertensive phenotypes. Methods: We included 10,221 individuals aged between 18-70 years from the multi-ethnic HELIUS study. Finger BP was recorded continuously for 3-5 minutes from which cross-correlation baroreflex sensitivity (xBRS) and heart rate variability (HRV) were determined. Hypertension was classified into isolated systolic (ISH; ≥140/<90), diastolic (IDH; <140/≥90) and combined systolic and diastolic hypertension (SDH; ≥140/≥90). Differences were assessed after stratification by age (younger: ≤40, older: >40 years) and sex, using regression with correction for relevant covariates. For xBRS, values were log-transformed. Results: In younger adults with ISH, xBRS was comparable to normotensive individuals in men (ratio 0.92; 95%CI 0.84-1.01) and women (1.00; 95%CI 0.84-1.20), while xBRS was significantly lower in IDH and SDH (ratios between 0.67 and 0.80). In older adults, all hypertensive phenotypes had significantly lower xBRS compared to normotensives. We found a similar pattern for HRV in men, while in women HRV did not differ between phenotypes. Conclusions: In younger men and women ISH is not associated with a shift towards increased sympathetic control, while IDH and SDH in younger and all hypertensive phenotypes in older participants were associated with increased sympathetic control. This suggests that alterations in autonomic regulation could be a contributing factor to known prognostic disparities between hypertensive phenotypes.

Chest pain is a common and challenging symptom for telephone triage in urgent primary care. Existing chest-pain-specific risk scores originally developed for diagnostic purposes may outperform current telephone triage protocols. This study involved a retrospective, observational cohort of consecutive patients evaluated for chest pain at a large-scale out-of-hours primary care facility in the Netherlands. We evaluated the performance of the Marburg Heart Score (MHS) and INTERCHEST score as stand-alone triage tools and compared them with the current decision support tool, the Netherlands Triage Standard (NTS). The outcomes of interest were: C‑statistics, calibration and diagnostic accuracy for optimised thresholds with major events as the reference standard. Major events are a composite of all-cause mortality and both cardiovascular and non-cardiovascular urgent underlying conditions occurring within 6 weeks of initial contact. We included 1433 patients, 57.6

Abstract Funding Acknowledgements Type of funding sources: None. Introduction New risk prediction models estimate and employ individual ‘treatment benefit’, which can be used to motivate patients with atherosclerotic cardiovascular disease (ASCVD) to quit smoking and to adhere to beneficial pharmacological interventions. However, this treatment benefit is usually calculated for a limited set of cardiovascular outcomes, i.e. years gained without myocardial infarction or stroke, while ignoring non-cardiovascular health benefits and pharmacological side- and adverse effects. Importantly, treatment effect size of medication is smaller in persistent smokers compared to non-smokers, because of the higher overall mortality of the smokers. By disregarding non-cardiovascular outcomes, the overall benefit of smoking cessation will be underestimated. Purpose We estimated and compared the treatment benefits – expressed as ‘gain in years without major cardiovascular events’ – of smoking cessation versus persistent smoking with targeted pharmaceutical interventions in patients with established ASCVD treated with anti-platelet agents, statins and anti-hypertensive drugs. Methods We pooled individual-level risk factors data from six large, recent prospective studies: RESPONSE 1 and 2, OPTICARE, EUROASPIRE IV and V and HELIUS. We included patients aged ≥45 years who persisted in smoking ≥6 months after acute coronary syndrome or revascularisation. The primary outcome was SMART-REACH estimated treatment benefit expressed as gain in years without a myocardial infarction or stroke. We compared the cardiovascular treatment benefit of smoking cessation versus the use of one or more pharmaceutical treatments: bempedoic acid, colchicine and PCSK9 inhibitors. Results We included 989 smokers with established ASCVD (23% female), with mean age of 60 (SD 8) years at median 1.2 (IQR 1.0-2.0) years post-index event. A mean of 4.81 (95%CI 4.73-4.89) event-free years would be gained through smoking cessation. Persistent smoking with maximal pharmaceutical treatment resulted in a comparable gain of 4.83 (95% CI 4.72-4.93) event-free years.(Figure) Conclusion The estimated lifetime treatment benefit of smoking cessation appeared to be comparable to the use of several pharmaceutical treatments combined, even when the analysis was limited to major cardiovascular events. This substantial health benefit underscores smoking cessation to be one of the most important actions to improve the overall health of patients with established ASCVD. To accurately compare treatment options, overall benefits and harms should be considered, in addition to the patients’ preferences, in a shared decision making process.

In old age, cognitive functions decline in comparison to performance earlier in life. Subjective memory complaints often cause anxiety. Patient and informant history taking, along with simple additional cognitive tests are usually sufficient to reassure patients or to initiate further diagnostic work-up. Differential diagnoses include depression, delirium, other psychiatric or neurological disorders, metabolic causes, intoxications or side effects of drugs. Physical examination and laboratory tests generally have limited added value. Main reasons for referral are a suspected treatable cause for cognitive decline (e.g. delirium), focal neurological deficit (e.g. stroke), suspected dementia under the age of 65-70 years, additional complex somatic or psychiatric disorders or further work-up in non-native speaking people.

Background: Women at risk of cardiovascular disease (CVD) may be missed with current eligibility criteria for CVD risk screening, particularly those from ethnic minority groups, among whom high risk is prevalent at a younger age. Early menopause (EM; menopause before 45 years) is associated with increased risk of CVD, and may be a potential eligibility criterion for CVD risk screening. Aims and objectives: To determine the contribution of EM to current criteria from patient history (having a family history of CVD, current smoking, obesity and age over 50 years) for identifying women eligible for CVD risk screening in a multi-ethnic population. Methods and results: We used baseline data (2011-2015) from 4512 women aged 45-70 years of Dutch, SouthAsian Surinamese, African Surinamese, Ghanaian, Turkish and Moroccan ethnic origin from the HELIUS study (Amsterdam, Netherlands). Models based on current eligibility criteria with and without EM were compared on area under the curve (AUC) with regard to estimated 10-year CVD risk using the Dutch SCORE. Overall, models with EM had a higher AUC, but changes were not statistically significant. In our total sample of women aged between 45 and 70 years, the AUC changed from 0.70 (95%CI 0.69-0.72) to 0.71 (95%CI 0.69-0.72). Among women aged 45-50 years the AUC changed from 0.66 (95%CI 0.58-0.74) to 0.68 (95%CI 0.59-0.74). Results were consistent across ethnic groups. Conclusions: The addition of EM to current eligibility criteria did not improve the detection of women at high CVD risk in a multi-ethnic sample of women aged 45-70 years.

Aims: We aimed to describe differences in the prevalence of intermediate hyperglycaemia (IH) between six ethnic groups. Moreover, to investigate differences in the association of the classifications of IH with the incidence of T2DM between ethnic groups. Methods: We included 3759 Dutch, 2826 African Surinamese, 1646 Ghanaian, 2571 Turkish, 2691 Moroccan and 1970 South Asian Surinamese origin participants of the HELIUS study. IH was measured by fasting plasma glucose (FPG) and HbA1c. We calculated age-, BMI and physical-activity-adjusted prevalence of IH by sex, and calculated age and sex-adjusted hazard ratios (HR)for the association between IH and T2DM in each ethnic group. Results: The prevalence of IH was higher among ethnic minority groups (68.6-41.7%) than the Dutch majority (34.9%). The prevalence of IH categories varied across subgroups. Combined increased FPG and HbA1c was most prevalent in South-Asian Surinamese men (27.6%, 95 %CI: 24.5-30.9%), and in Dutch women (4.2%, 95 %CI: 3.4-5.1%). The HRs for T2DM for each IH-classification did not differ significantly between ethnic groups. HRs were highest for the combined classification, e.g., HR = 8.1, 95 %CI: 2.5-26.6 in the Dutch. Conclusion: We found a higher prevalence of IH in ethnic minority versus majority groups, but did not find evidence for a differential association of IH with incident T2DM.

Abstract Funding Acknowledgements Type of funding sources: None. Introduction Most risk models for patients with established atherosclerotic cardiovascular disease (ASCVD) calculate short-term risk of recurrent events and death, typically for a duration of 10 years. However, lifetime risk estimates may better support the healthcare professional in selecting patients for intensified preventive treatment (1). Also, a cross-sectional study suggested that communicating lifetime risk to ASCVD patients enhances risk perception and willingness for therapy (2). In the new ESC prevention guideline, however, 10-year risk estimates remain standard for ASCVD patients but the additional use of lifetime risk is recommended for communication in the shared decision-making process (3). Purpose We therefore aimed to compare estimates of 10-year with lifetime risk of recurrent ASCVD events or death, stratified by age. Methods We pooled individual-level data on risk factors from six large, recent prospective studies (RESPONSE 1 and 2, OPTICARE, EUROASPIRE IV and V and HELIUS). We included Dutch patients aged ≥45 years with a follow-up of ≥6 months after acute coronary syndrome or revascularisation. The SMART-REACH models were used to estimate the difference between 10-year and lifetime risk of recurrent myocardial infarction, stroke, or cardiovascular death, stratified by age (<55, 55-65, 65-75, ≥75 years). Results In 3,230 ASCVD patients (24% women), mean age 61±8 years, at median follow-up 1.1 (IQR 1.0-1.8) years after index event, SMART-REACH 10-year risk was 23±11% versus lifetime 56±11%. (Figure 1) We found a considerable difference between 10-year and lifetime risk in patients aged 45-55 years (18±8% vs. 61±10%). Discrepancies decreased with increasing age, with similar estimates in the highest (75-85) age group. (Figure 2). Conclusion Lifetime risk of a limited set of cardiovascular outcomes rather than 10-year risk may provide a more complete estimate of future ASCVD disease burden, as especially in younger patients 10-year risk is usually low, even in the presence of risk factors.

Objectives Prevention and lifestyle support are emerging topics in general practice. Healthcare insurance companies reimburse combined lifestyle interventions (CLIs) in the Netherlands since January 2019. CLIs support people with overweight (body mass index, BMI 25-30) or obesity (BMI >30) to reduce weight in peer groups. General practitioners (GPs) are key in the successful implementation of lifestyle interventions in primary care. This study explored GPs' experiences and views on the implementation of CLIs to identify barriers and facilitators to the successful implementation in primary care. Design Qualitative study using semistructured interviews. Content analysis consisted of thematic coding and mapping a first stage of predefined and second stage of iterative evolving set of themes. Setting GPs were interviewed in a variety of primary care practices between February and April 2019. Participants Fifteen GPs were purposively recruited for semi-structured interviews through snowballing. Results Experiences with lifestyle support among GPs ranged from referring patients to other healthcare professionals to taking a proactive role in lifestyle support themselves. Whether or not GPs took an active role in lifestyle support was related to their belief in the effect of lifestyle interventions. Overall, GPs had little experience with CLI in every day practice. Perceived barriers were a lack of availability of CLIs in the region and the potential lack of added value of CLIs on top of existing lifestyle support. Perceived facilitators were coordination of care provision by GP cooperatives and monitoring of the CLI implementation and their results. Reimbursement of CLIs without any costs for participants enabled application. Conclusion The importance of lifestyle interventions in primary care was acknowledged by all GPs, but they differed in their level of experience with providing lifestyle support and awareness of CLIs. Successful integration of CLIs with primary care requires a solid promotion, a well-coordinated implementation strategy and structural evaluation of long-term effectiveness.

Abstract Background Cardiovascular diseases (CVD) are a major public health problem in men and women worldwide. Differences in CVD incidence between men and women have been widely reported, although explanations for these differences remain unclear. Next to sex-related (biological) characteristics, gender-related (sociocultural) characteristics may partly explain how these differences arise. Purpose We aimed to establish the associations between selected gender-related characteristics and CVD incidence. Methods We linked baseline data of 18,058 participants without CVD from the multi-ethnic HELIUS study (Amsterdam, the Netherlands) to CVD incidence data, based on hospital admission and death records from Statistics Netherlands in 2013–2018. Using Cox regression analyses, we studied associations of time spent on household work, doing home repairs, primary earner status, type of employment, performing a male-dominated or female-dominated occupation, and desire for emotional support with CVD incidence, overall (with or without adjustment for sex) and stratified by sex. Analyses were adjusted for age, ethnicity, and socioeconomic status. Results In the fully adjusted analyses, most gender-related characteristics were not associated with higher CVD incidence, overall and among men. In women, homemakers had a higher hazard for CVD compared to fulltime workers (hazard ratio [HR] 2.34, 95% confidence interval [CI] 1.35–4.04), whereas those spending a moderate amount of time on household work had a lower hazard for CVD than those spending little time (HR 0.56, 95% CI 0.34–0.95). Conclusion Although most gender-related characteristics were not associated with CVD incidence in men, being the homemaker and time spent on household work were associated with CVD incidence in women. Thus, attention to gender-related characteristics may in future help to identify subgroups that may benefit from additional prevention strategies, particularly in women. Funding Acknowledgement Type of funding sources: Public grant(s) – National budget only. Main funding source(s): ZonMw Gender and Health (grant number 8492ehab724.26598) and Gender and Prevention (grant number 555003016) programmes.European Union's Horizon 2020 research and innovation programme under acronym ESCAPE-NET (grant number 733381) and the COST Action PARQ (grant agreement number CA19137).

Abstract Introduction The major modifiable risk factors for atherosclerosis – lifestyle, hypertension, diabetes and cholesterol – collectively account for 80 to 90% of disease burden. Currently, the majority of coronary patients does not meet the guideline-directed treatment targets for these risk factors, resulting in high levels of residual risk. An increasing number of novel preventive drugs aims to reduce this residual risk, but are not considered cost-effective when added routinely to all patients. Quantifying the potential lifetime risk reduction one year after an acute coronary syndrome (ACS) may aid in optimum use of available treatment and value-based use of novel drugs. Purpose The purpose of this analysis was to quantify the loss of lifetime risk reduction due to suboptimal modifiable risk factor control in patients with prior ACS or revascularisation. Methods We pooled six recent prospective studies (Response 1 [1] and 2 [2], Opticare [3], EuroAspire IV [4] and V [5] and HELIUS [6]) with Dutch patients (n=3,230, 24% women) at mean age 61±8 years and follow-up at median 1.1 [IQR 1.0–1.8] years after an ACS or revascularisation. We investigated individual lifestyle- and drug-modifiable risk factors at guideline-directed targets. Using the SMART-REACH model [7], we calculated % reduction of individual residual lifetime risk for myocardial infarction, stroke, or cardiovascular death and event free years gained by the change from current treatment to a (simulated) guideline-directed optimal situation. Results Risk factor control was far from optimal: only 7% met all lifestyle-related risk targets, whereas 10% met none: 30% persist smoking, 79% was overweight (BMI ≥25 kg/m2), of which 40% obese (BMI ≥30 kg/m2), and 45% reported insufficient physical activity (<150 minutes per week). Systolic blood pressure ≥140 mmHg was found in 40%, and LDL-cholesterol ≥1.8 mmol/L or ≥2.5 mmol/L (depending on the target at that time) in 65%. Basic preventive medication use was, however, common: 87% used antithrombotic agents, 85% lipid lowering drugs and 86% any blood pressure lowering drugs. By the change from current to optimal guideline-directed treatment, residual lifetime risk for cardiovascular events and cardiovascular death would decrease from a mean of 54±11% to 25±10% (Figure 1), and a median of 7.4 [IQR 5.2–10.6] event free years would be gained (Figure 2). Conclusion Suboptimal risk factor control resulted in avoidable high residual lifetime risk of myocardial infarction, stroke, or cardiovascular death and loss of event free years in patients with prior ACS or revascularisation. This finding highlights the unexploited potential of optimised use of available lifestyle- and drug treatment to significantly reduce residual lifetime risk. Funding Acknowledgement Type of funding sources: None. Figure 1. Modifiable residual lifetime riskFigure 2. Lifetime benefit in CVD event free years

Introduction Profiles of high risk for future dementia are well understood and are likely to concern mostly those in low-income and middle-income countries and people at greater disadvantage in high-income countries. Approximately 30%–40% of dementia cases have been estimated to be attributed to modifiable risk factors, including hypertension, smoking and sedentary lifestyle. Tailored interventions targeting these risk factors can potentially prevent or delay the onset of dementia. Mobile health (mHealth) improves accessibility of such prevention strategies in hard-to-reach populations while at the same time tailoring such approaches. In the current study, we will investigate the effectiveness and implementation of a coach-supported mHealth intervention, targeting dementia risk factors, to reduce dementia risk. Methods and analysis The prevention of dementia using mobile phone applications (PRODEMOS) randomised controlled trial will follow an effectiveness–implementation hybrid design, taking place in the UK and China. People are eligible if they are 55–75 years old, of low socioeconomic status (UK) or from the general population (China); have ≥2 dementia risk factors; and own a smartphone. 2400 participants will be randomised to either a coach-supported, interactive mHealth platform, facilitating self-management of dementia risk factors, or a static control platform. The intervention and follow-up period will be 18 months. The primary effectiveness outcome is change in the previously validated Cardiovascular Risk Factors, Ageing and Incidence of Dementia dementia risk score. The main secondary outcomes include improvement of individual risk factors and cost-effectiveness. Implementation outcomes include acceptability, adoption, feasibility and sustainability of the intervention. Ethics and dissemination The PRODEMOS trial is sponsored in the UK by the University of Cambridge and is granted ethical approval by the London—Brighton and Sussex Research Ethics Committee (reference: 20/LO/01440). In China, the trial is approved by the medical ethics committees of Capital Medical University, Beijing Tiantan Hospital, Beijing Geriatric Hospital, Chinese People’s Liberation Army General Hospital, Taishan Medical University and Xuanwu Hospital. Results will be published in a peer-reviewed journal. Trial registration number ISRCTN15986016.

Objective To determine variation in diagnostic strategies for diagnosing dementia between Dutch hospitals. Design Descriptive, retrospective research based on claim data of Dutch health insurers. Method Information on the use of diagnostic ancillary services carried out from 2015 to 2018 was collected via national-level insurance claims for patients who received a (new) diagnose-coding for dementia in 2018. Hospitals were included in the analysis if they diagnosed >50 patients with dementia. We distinguished academic medical centres (AMC), non-academic training hospitals (TH) and general hospitals (GH). Results In 2018, 20.073 new cases of dementia were diagnosed in 71 hospitals. The percentages of patients undergoing MRI/CT-imaging ranged from 37 to 99% (median 76.7%), neuropsychological-assessment from 0-89% (median 31.8%), cerebrospinal fluid examination from 0-14% (median 2.4%), PET/SPECT-imaging from 0-16% (median 6.2%) and electroencephalography from 1-20% (median 5.8%). Practice variation was comparable in AMCs, THs and GHs and was evidently skewed for PET/SPECT-imaging, electroencephalography and cerebrospinal fluid examination. There were no distinct differences according to case-mix characteristics or hospital volume. The percentage of patients subjected to ancillary diagnostic investigations decreased sharply with increasing age. Conclusion In the Netherlands, diagnostic ancillary methods used vary widely between hospitals both in frequency and modality. This variation may be driven by limited evidence of diagnostic accuracy and added value of different diagnostic tests, variations in doctor and patient preferences and differences in available diagnostic techniques per hospital. Further exploration of this heterogeneity may help to identify a strategy that combines the most benefit with the least burden.

Objective Differences in cardiovascular disease (CVD) incidence between men and women have been widely reported. Next to sex-related (biological) characteristics, gender-related (sociocultural) characteristics may partly explain how these differences arise. In this exploratory study, we examined the associations between selected gender-related characteristics and CVD incidence. Methods We linked baseline data of 18 058 participants without CVD from the population-based, multiethnic HEalthy LIfe in an Urban Setting study (Amsterdam, the Netherlands) to CVD incidence data, based on hospital admission and death records from Statistics Netherlands in 2013-2018. Using Cox regression analyses, we studied associations of time spent on household work, doing home repairs, primary earner status, type of employment, working in a male-dominated or female-dominated occupation and desire for emotional support with CVD incidence, stratified by sex. Analyses were adjusted for age, ethnicity and socioeconomic status. Results In men, gender-related characteristics were not associated with higher CVD incidence. In women, homemakers had a higher hazard for CVD compared with full-time workers (HR 2.34, 95% CI 1.35 to 4.04), whereas those spending a moderate amount of time on household work had a lower hazard for CVD than those spending little time (HR 0.56, 95% CI 0.34 to 0.95). Conclusion Although we found no evidence for associations between gender-related characteristics and CVD incidence in men, being the homemaker and moderate time spent on household work appeared to be associated with CVD incidence in women. Thus, attention to gender-related characteristics might in future help to identify subgroups that may benefit from additional prevention strategies.

Abstract Background Among patients using antihypertensives, blood pressure (BP) control is generally low and appears to vary between ethnicities. We aimed to identify factors associated with suboptimal BP control in a multi-ethnic population. Methods Cross-sectional data of the Healthy Life in an Urban Setting (HELIUS) study were used, including participants aged 18-70 years, with Dutch, African Surinamese (AS), South-Asian Surinamese (SAS), Ghanaian, Turkish or Moroccan ethnicity living in Amsterdam, the Netherlands. Suboptimal BP control was defined as BP ≥ 140/90 mmHg among participants using antihypertensives. Logistic regression analysis was performed to identify sociodemographic, lifestyle, health and psychosocial factors associated with suboptimal BP control for the total population and for each ethnic group. Results A total of 3,571 participants (500 Dutch, 1052 AS, 656 SAS, 637 Ghanaian, 433 Turkish, 293 Moroccan) were included in the analysis, of whom 53.3% had suboptimal BP control. Female sex (OR 0.50, 95%CI 0.43-0.59), being married (0.83, 0.72-0.96), current smoking (0.78, 0.65-0.94), obesity (1.67, 1.35-2.06), cardiovascular disease (CVD) history (0.56, 0.46-0.68), non-adherence to antihypertensives (1.26, 1.00-1.58), and family history of hypertension (1.19, 1.02-1.38) were associated with suboptimal BP control. Among ethnic subgroups, female sex was associated with lower odds of suboptimal BP control in all ethnic groups, as was current smoking (Turkish) and CVD history (Dutch, SAS, AS), whereas obesity (Dutch, AS, Turkish), older age (Turkish) and non-adherence to antihypertensives (Dutch) were associated with higher odds of suboptimal BP control. Conclusions Our analysis identifies several factors that are independently associated with suboptimal BP control, with only small variation between ethnic groups, suggesting that focussing on well-known determinants of suboptimal BP control is worthwhile to improve BP control rates in a multi-ethnic population. Key messages Several factors are associated with suboptimal blood pressure control in a multi-ethnic population, with only small variation between ethnic groups. These findings suggest that focussing on well-known determinants of suboptimal BP control is worthwhile to improve BP control rates in a multi-ethnic population.

With increasing age, associations between traditional risk factors (TRFs) and cardiovascular disease (CVD) shift. It is unknown which mid-life risk factors remain relevant predictors for CVD in older people. We systematically searched PubMed and EMBASE on August 16th 2019 for studies assessing predictive ability of >1 of fourteen TRFs for fatal and non-fatal CVD, in the general population aged 60+. We included 12 studies, comprising 11 unique cohorts. TRF were evaluated in 2 to 11 cohorts, and retained in 0-70% of the cohorts: age (70%), diabetes (64%), male sex (57%), systolic blood pressure (SBP) (50%), smoking (36%), high-density lipoprotein cholesterol (HDL) (33%), left ventricular hypertrophy (LVH) (33%), total cholesterol (22%), diastolic blood pressure (20%), antihypertensive medication use (AHM) (20%), body mass index (BMI) (0%), hypertension (0%), low-density lipoprotein cholesterol (0%). In studies with low to moderate risk of bias, systolic blood pressure (SBP) (80%), smoking (80%) and HDL cholesterol (60%) were more often retained. Model performance was moderate with C-statistics ranging from 0.61 to 0.77. Compared to middle-aged adults, in people aged 60+ different risk factors predict CVD and current prediction models perform only moderate at best. According to most studies, age, sex and diabetes seem valuable predictors of CVD in old-age. SBP, HDL cholesterol and smoking may also have predictive value. Other blood pressure and cholesterol related variables, BMI, and LVH seem of very limited or no additional value. Without competing risk analysis, predictors are overestimated.